Conventional motion detection looks at a local error measure, commonly a sum-of-absolute-differences, between a target picture and a reference picture. Even if no motion exists, such local error measures tend to be non-zero due to noise and changes in scene lightness. Therefore, motion detection commonly detects small differences between the pictures as no motion and detects big differences as motion. Temporal filtering is used to combine a target picture with a motion compensated reference picture, and uses strong filtering where no motion is detected.
It would be desirable to implement motion detection based on observing several pictures.